Quote from Jorgen Bergström :
User-subroutines were recently published on the SIMULIA Knowledge-Base to allow the Abaqus PRF model to replicate Bergström's Three Network Model. Let us call this the "PRF_TNM" model. Abaqus solver 2025 HF1 or later can be used. This post describes how the 3DExperience Material Calibration App can be used to calibrate the PRF_TNM model. I will be showing this with the development version of the app, and this capability will be released as of R2026x FD01.
For this demonstration we will use test data for UHMWPE (Ultra High Molecular Weight Polyethylene) digitized by Victor Lefevre from the 2010 Bergström paper [1]. This digitized test data is attached in the zip file at the end of this post.
- Data from figures 1a, 1b, 2b, 4a, and 4b from Bergström, J., Bischoff, J., 2010.
- The rates indicated on those figures of the original paper are incorrect. The correct rates can be found in this YouTube video https://www.youtube.com/watch?v=vJ45UWwWido from the PolymerFEM channel.
There are 7 uniaxial tests which are homogenous and can be used in Numerical mode. There is 1 test, the hemispherical punch test, which must use the FE mode.
Part A. Calibration in Numerical mode using 7 uniaxial test datasets.
If you were to import & open the file UHMWPE_True_B&B_2010_just_data.3dxml it would look like this:
In the video below I will start here and show the process of selecting / creating the PRF_TNM material model. This will include:
Selecting the PRF material model
Select UHYPER as the elastic part.
Select UCREEPNETWORK as the viscous network part.
Select to use solution dependent variables (DEPVAR)
Select to use and import the Custom Schema. This will define all the parameters, with names, initial values, and units. The initial parameter values are the reference values converted from the 2010 paper.
Import the standardU.dll for the two user-subroutines (compiled using Abaqus make).
This is an important part of Victor Lefevre's PDF document in the KB article:
Note: In Jorgen's calibration tool, MCalibration, there is a special feature that allows the user to set a target Poisson's ratio. In YouTube videos, Jorgen set the Poisson's ratio = 0.4 for thermoplastics. The value of Kappa (bulk modulus) could be defined as function of "mu_inst" and an a target Poisson’s ratio using:
The is no test data in this example that allows us to determine Kappa.
Part B. Calibration in FE mode using all 8 test datasets.
In order to also use the hemispherical punch test result in the calibration, we would have to use FE mode. This means that we would have to convert the 7 Numerical mode test datasets into unit-cube FE models. There is a new feature in R2026x FD01 to help you do this. The video below shows how to create the 7 unit cube FE models.
Another video showing the use of FE mode will be inserted here soon
The zip file below contains: (coming soon)
The 2010 reference paper.
The test data that Victor Lefevre digitized from the 2010 paper.
All of the files from the Knowledge-Base article (includes the Fortran source code files).
The Custom Schema .xml file
An Excel file that converts from Jorgen's TNM parameters to the required PRF_TNM parameters.
Reference:
Bergström, J., Bischoff, J., 2010. An advanced thermomechanical constitutive model for UHMWPE. The International Journal of Structural Changes in Solids, 2 (1), 31–39.
